Download Pied Piping and the Word Orders of English and Dutch Jan Koster

Pied Piping and the Word Orders of English and Dutch
Jan Koster
University of Groningen
1.
Introduction
A remarkable aspect of the word order of Dutch is that, in V-final non-root clauses, the
verb serves as a "mirror center" for series of PPs, i.e., the unmarked order of PPs to the
left of the final verb is mirrored on the right (Barbiers 1995, Koster 1974):
(1)
(2)
a.
b.
Hij heeft tijdens de pauze aan zijn vader gedacht
he has during the break of his father thought
"He thought of his father during the break"
*Hij heeft aan zijn vader tijdens de pauze gedacht
a.
b.
Hij heeft gedacht aan zijn vader tijdens de pauze
*Hij heeft gedacht tijdens de pauze aan zijn vader
Mirror symmetry is caused by parallel construal (my alternative for extraposition; see
Koster 1999b) and is "broken" in root clauses, which was one of the main arguments for
the verb movement rule known as Verb Second (see Koster 1975).
In English, we find only one possible order, namely the Dutch order (2a):
(3)
a.
b.
He thought of his father during the break
*He thought during the break of his father
As in Dutch root clauses, there is no mirror symmetry in English here, since the PPs
cannot appear to the left of the verb (apart from topicalization):
© 2000 by Jan Koster
NELS 30
2
(4)
Koster
*He during the break of his father thought
Under the antisymmetry theory of Kayne (1994), which is adopted here, these facts are
puzzling because the surface word orders of Dutch and English must be derived from the
same underlying SVO order. Supposing that the mirror image orders observed in (1) are
common to both English and Dutch at some level, the problem to be solved is: how is the
underlying mirror symmetry broken in English?
The answer to the same question given for Dutch, the postulation of a rule of Verb
Second, does not work for English, as is well-known: English is simply not a verb second
language, as all kinds of material can intervene between the first constituent and the finite
verb in root clauses. In this respect, English (5a) is in sharp contrast with Dutch (5b):
(5)
a.
b.
Peter probably saw Bill
*Peter waarschijnlijk zag Wim
Peter probably
saw Bill
The solution to the observed problems can be based on recent analyses of the structure of
Dutch. The key idea goes back to Vanden Wyngaerd's (1989) insight that the Dutch
object in OV structures is not in its base position, but in a derived position in order to
check its case features. This position was identified as the Spec of AgrOP by Vanden
Wyngaerd, i.e., a position in the functional shell containing the VP.
As was first concluded by Jan-Wouter Zwart, Vanden Wyngaerd's rule makes it
possible to derive Dutch OV orders from Kayne's universal underlying VO order. This
derivation, apart from bringing Dutch more in tune with universal grammar, happened to
have a number of empirical advantages (see Zwart 1993, and also Kaan 1992, Koster
1994 and Den Dikken 1996).
An immediate problem arising in the analyses in question is that, at first sight,
English does not show overt object shift for case checking at all. This led to the arbitrary
and therefore unsatisfactory idea that English case features are "weak" (as opposed to the
"strong" features of Dutch) and checked by covert movement at LF.
In this paper, I will show that English checking for case features is just as overt as
in Dutch, but that instead of the object, the whole VP containing the object is "moved" to
the checking position. In other words, English overt movement involves Pied Piping of
the whole VP (see Koopman and Szabolcsi 1998 for similar cases of "massive" Pied
Piping).
Recall that the mirror symmetry of PPs with respect to the verb was broken in
Dutch by movement of the verb. The Pied Piping solution for English case checking
explains the 25-year old problem why the symmetry is broken in English, too. As in
Dutch, the underlying pattern is blurred by verb movement, but in English the verb is
moved as part of the entire VP. Apart from this fundamental problem, many other word
order puzzles about English and Dutch are solved as well (Koster 1999a).
2.
The Facts
Comparing English with Dutch (and implicitly also with German) we observe the
following facts:
The Word Orders of English and Dutch
(6)
a.
b.
c.
d.
e.
3
English is VO, Dutch is OV
Neither English nor Dutch has rightward scrambling
Unlike Dutch, English has no leftward scrambling
In Dutch all Adv's can appear to the left of the VP, in English only a
subclass
English Adv order shows a scope paradox (absent from Dutch)
In the remainder of this section, I will illustrate these classes of facts one by one.
2.1
English is VO, Dutch is OV
If we compare English and Dutch non-root clauses, we observe that the object is to the
right of the verb in English (7a) and to the left in Dutch (7b). The English order is sharply
ungrammatical in Dutch (7c):
(7)
a.
b.
c.
2.2
that John read the book
dat Jan het boek las
that John the book read
*dat Jan las het boek
Neither English nor Dutch has Rightward Scrambling
Before Kayne (1994), theories of Universal Grammar often did not exclude adjunction of
APs or NPs to the right of the sentence. This makes the following elementary facts of
English very puzzling:
(8)
a.
b.
John read the book yesterday
*John read yesterday the book
(9)
a.
b.
John gave Bill a book
*John gave a book Bill
(10)
a.
b.
Mary made Sue happy
*Mary made happy Sue
(11)
a.
b.
Mary was happy yesterday
*Mary was yesterday happy
It is simply impossible to adjoin NPs or APs to the right and exactly the same can be
observed in Dutch:
(12)
a.
b.
Jan heeft het boek gelezen
John has the book read
"John has read the book"
*Jan heeft gelezen het boek
John has
read the book
4
(13)
Jan Koster
a.
b.
(14)
a.
b.
(15)
a.
b.
2.3
Jan heeft Mary een boek gegeven
John has Mary a book given
"John has given Mary a book"
*Jan heeft een boek gegeven Mary
John has a book given Mary
Mary heeft Suus gelukkig gemaakt
Mary has Sue happy made
"Mary has made Sue happy"
*Mary heeft Suus gemaakt gelukkig
Mary has Sue
made happy
Mary is gelukkig geweest
Mary is happy been
"Mary has been happy"
*Mary is geweest gelukkig
Mary is been happy
Unlike Dutch, English Has No Leftward Scrambling
Although English and Dutch behave exactly the same with respect to rightward
adjunction, we see a very striking difference with respect to movement to the left
("scrambling"). In English, apart from Wh-movement and related rules, it is impossible
to move VP-internal material to the left around VP-external adverbials such as probably:
(16)
*that John the book probably read t
In Dutch, however, such forms of scrambling lead to very natural word orders:
(17)
dat Jan het boek waarschijnlijk las
that John the book probably
read
"John probably read the book"
In fact, the English VP behaves like a cage for it constituents: apart from frontings such
as Wh-movement, the VP-internal constituents (except the non-V-specific subject) can be
moved neither to the right nor to the left. The VP is closed on both sides. Dutch and
German, in contrast, only seem closed to the right.
2.4
In Dutch All Adv's Can Appear to the Left of the VP, in English Only a
Subclass
A very curious fact of English is that only a subclass of the adverbials can appear to the
left of the VP. Thus, probably can appear to the left of the VP, while yesterday etc.
cannot. Under a theory which assumes that adverbials are freely adjoined to VPs, these
facts are totally unexpected:
The Word Orders of English and Dutch
(18)
(19)
(20)
(21)
5
He probably [VP saw Bill]
*He yesterday [VP saw Bill]
*He everywhere [VP saw Mary]
*He very hard [VP worked]
In Dutch, in contrast, all adverbials in question can appear to the left of what is
traditionally considered the VP (actually a bigger constituent such as AgrOP):
(22)
dat hij waarschijnlijk [VP Wim zag]
that he probably
Bill saw
"that he probably saw Bill"
(23)
dat hij gisteren [VP Wim zag]
that he yesterday
Bill saw
"He has saw Bill yesterday"
(24)
dat hij overal [VP Mary zag]
that he everywhere Mary saw
"He has seen Mary everywhere"
(25)
dat hij erg hard [VP werkte]
that he very hard
worked
"He worked very hard"
2.5
English Adv Order Shows a Scope Paradox (Absent from Dutch)
Very often, the scope order of adverbials is linear in both English and Dutch, i.e. the
wider the scope of an adverbial the more it appears to the left. Thus, in the following
Dutch example, twee keer ("twice") has wider scope than op zijn verjaardag (“on his
birthday”) to its right:
(26)
Hij heeft Wim twee keer op zijn verjaardag gezien
he has Bill twice
on his birthday seen
"He saw Bill twice on his birthday"
If we reverse the linear order of the adverbials, the scope can be reversed as well:
(27)
Hij heeft Wim op zijn verjaardag twee keer gezien
he has Bill on his birthday
twice
seen
"He saw Bill on his birthday twice "
These facts are unproblematic under the traditional assumptions that adverbials are
successively adjoined to the left of the VP, i.e. "being in a higher, c-commanding
position" corresponds with linear order:
(28)
[Adv [Adv [VP]]]
6
Jan Koster
Also, according to traditional assumptions, nothing much changes when adverbial
material is VP-internal, like hard ("hard") in (29):
(29)
dat hij [gisteren [VP hard werkte]]
that he yesterday hard worked
"that he worked hard yesterday"
As before, the adverbial with the wider scope, gisteren, precedes the VP-internal
adverbial hard, which has the narrower scope. So, a reasonable assumption is that
adverbial scope is always linear in underlying structure, i.e., the wider an adverbial's
scope, the more it is to the left in the structure.
The fact that English is extremely irregular from this point of view has not
received the attention it deserves. Just consider the facts corresponding to the Dutch
examples (">" means: "has wider scope than"):
(30)
a.
b.
He saw Bill twice on his birthday [twice > on his birthday]
He saw Bill on his birthday twice [on his birthday > twice]
(I will here ignore the fact that the opposite scope order is also possible, thanks to the
mechanisms underlying the mirror effect mentioned at the beginning of this article). Until
recently, it was very often assumed that one can add adverbials to the right of the VP by
adjoining them successively, yielding the opposite from the Dutch pattern illustrated by
(28)):
(31)
[[[VP] Adv] Adv]
This would be entirely anomalous, because usually "being in a higher c-commanding
position" means wider scope. So, if (31) were right, we would expect the rightmost
adverbial to always have the widest scope. This is contrary to fact, as illustrated by (30):
the scope facts can be exactly the same as in Dutch, i.e., the rightmost adverbial can have
the narrower scope rather than the wider scope. If we look at VP-internal adverbials,
however, we find exactly the opposite pattern:
(32)
He [[VP worked hard] yesterday]
For this case, the traditional assumption (rightward adjunction) would work, because
yesterday --the element with the wider scope-- would be in a c-commanding position
higher than hard in (32).
This situation is paradoxical because, proceeding from left to right, scope
sometimes becomes narrower (as in (30)) and sometimes wider (as in (32)). In fact, we
find the Dutch pattern (28) in English, except when the adverbial is VP-internal (as in
(32)). In the latter case, the Dutch order is in fact impossible:
(33)
*He worked yesterday hard
The Word Orders of English and Dutch
7
The same is true for other cases in which traditionally the adverbial is analyzed as part of
the VP (Jackendoff 1972, 64):
(34)
a.
b.
Steve dressed elegantly yesterday
*Steve dressed yesterday elegantly
The scope paradox of English adverbials makes a uniform solution along traditional lines
(Adv's going up) or Larsonian lines (Adv's going down) very unlikely. What the facts
strongly suggest is that English preserves the structural patterns of Dutch and German
except for the VP and its constituents. At least in this domain of facts, English looks like
Dutch, but with a VP that is displaced somehow.
3.
A Proposal
The hypothesis I would like to suggest as an explanation of the facts observed is based on
the idea that English has a rule of VP-movement. The theories I assume as background
are a form of Minimalism (Chomsky 1995), antisymmetry theory (Kayne 1994) and the
theory of the configurational matrix (Koster 1987, 1999a). In accordance with the latter
two theories, I assume that all languages are underlyingly head-initial, which means a VO
structure at the deepest level for both English and Dutch.
I also assume a theory of adverbial positions along the lines of Alexiadou (1997)
and Cinque (1998). Altogether, the theories assumed come down to the position that the
universal structure for all languages is [Spec [Head Complement]]. The head can be
lexical or functional and in general, lexical projections are embedded in a shell of
functional projections to their left.
As for these functional projections, I assume the conventional heads AgrS and T
(Tense). For the checking of case features, I further assume an Acc head for the
accusative and a Dat head for the Dative, which means that I reinterpret Vanden
Wyngaerd's object shift as a rule which moves the direct object to the Spec of the AccP
rather than the Spec of AgrOP. This preference is motivated by the fact that Dative DPs
must be moved to the left of the verb as well, so that one position for object shift (the
Spec of AgrOP) is not enough.
The three resulting case checking positions (the Specs of AgrSP, DatP and AccP)
correspond with the three semantically unrestricted argument positions of Relational
Grammar (see, for instance, Perlmutter 1983).
Apart from these argument positions, I also assume a Pred Phrase for oblique
arguments and the secondary predicates as found in Small Clauses (see Zwart 1993 and
Koster 1994 for arguments). Altogether, the relevant part of the universal base structure
looks as follows (where the XPs indicate Spec-positions):
(35)
Universal Base Structure
...[XP AgrS [XP Adv1* [XP T [XP Dat [XP Acc [XP Adv2* [XP [Pred* [VP] ]]]]]]]]
For reasons of space, I have omitted the brackets between Specs and heads here. A star
(*) means that there can be more than one projection with the type of head in question.
8
Jan Koster
For adverbials I make the simplifying assumption that class Adv1 contains
adverbials like probably (of the kind that can precede the VP in English) and that class
Adv2 also contains adverbials like yesterday (of the kind that cannot precede the VP in
English).
Crucially, I am assuming that all VP-internal material must be licensed in some
functional projection to the left of the VP by overt movement. In practice, this means that
VP-internal material can be functionally licensed either in one of the independent
argument positions or in the Pred Phrase. The independent argument positions are the
familiar case positions for subject, indirect object and direct object in, respectively, the
Specs of AgrSP, DatP and AccP. All other VP-internal material, such as prepositional
objects, oblique objects and the predicate part of Small Clauses is licensed in the Spec of
some PredP (see Koster 1994 for details).
The simple parametric difference responsible for the word order difference
between English and Dutch (as summarized in section 2.) has to do with the size of the
checking phrase (Pied Piping). This is a known and uncontroversial dimension of
language variation. Thus, in English, Wh-movement can either involve a minimal Whphrase (36a) or a PP containing it (36b):
(36)
a.
b.
Who did you talk [PP with t ]?
[PP With whom] did you talk t ?
In Dutch, Pied Piping of the whole PP is obligatory in such cases:
(37)
a.
b.
*Wie heb je [PP met t ] gepraat?
who have you with
talked
[PP Met wie] heb je t gepraat?
with whom have you talked
As for the possible size of the checking phrase, there is quite a bit of variation among the
languages of the world, and in several cases even whole clauses can be pied piped (see,
for instance, Van Riemsdijk 1994).
Given this uncontroversial (but hardly understood) dimension of language
variation, the parametric difference between English and Dutch can be formulated in
terms of a simple difference in the size of the checking phrase for VP-specific material
(i.e., the V, its Tense and the elements V is subcategorized for, which excludes the
subject):
(38)
Pied Piping parameter distinguishing English from Dutch
Dutch checks its VP-internal constituents individually (by moving them
separately), English collectively (by moving the whole VP).
Apart from this minimal difference, checking is done by overt movement to the same
checking positions in the same universal structure (35). I assume that the checking
process of VP-specific material (which excludes the non-VP-specific subject) stops
somewhere at the Spec of TP (with some problems of detail requiring further research):
The Word Orders of English and Dutch
(39)
9
Collective Checking (VP-movement) in English:
...[XP AgrS [XP Adv1* [VP T [VP Dat [VP Acc [XP Adv2* [VP [Pred* [VP] ]]]]]]]]
The VP can be moved through all Spec positions as indicated. This makes it possible, via
Pied Piping, to check not only the case features for dative and accusative DPs, but also
the elements (such as the predicates of Small Clauses) that can be seen not as independent
arguments but only as part of the predicate of the sentence.
In Dutch (and German), in contrast, all the Spec positions "visited" by VP in (39)
are filled for checking by the individual VP constituents (such as DPs and APs, which can
be assumed to be included in the VPs moved in (39)):
(40)
Individual Checking in Dutch and German:
...[XP AgrS [XP Adv1* [XP T [DP Dat [DP Acc [XP Adv2* [AP [Pred* [VP] ]]]]]]]]
This simple difference between English and Dutch (and German) explains all major word
order differences as discussed in section 2.
4.
The Facts Explained
I will now show that the facts are actually explained by the Pied Piping parameter (and
the associated theoretical framework). Consider the VO/OV difference between English
and Dutch. For a long time, this difference was accounted for by an entirely arbitrary
parameter known as the VO/OV parameter. This parameter is no longer necessary.
English moves the whole VP to the case checking positions for the objects, thereby
leaving the original universal base order (VO) intact. Dutch and German, in contrast,
modify the relative order of objects and verb by moving the object DPs individually to the
checking positions, as indicated in (40).
The second word order fact, that neither English nor Dutch and German show any
evidence for rightward movement of VP-internal material just follows from antisymmetry
theory as formulated by Kayne (1994). This theory simply excludes rightward adjunction
of VP-internal material.
The fact that Dutch and German have leftward scrambling while English does not,
also follows immediately from the Pied Piping parameter. Leftward scrambling in Dutch
and German is just the manifestation of the individual movements of the VP-internal DPs
to the checking positions, which can be surrounded by all kinds of adverbial positions
(not all mentioned in (35) and (40)). In English, in contrast, the whole VP is moved up to
the Spec of TP, which does not leave any functional checking positions to the left of the
VP (apart from the Spec of AgrSP for the subject). This makes it impossible for English
to derive the same scrambling phenomena as found in Dutch and German.
The fourth class of facts concerned the absence of certain adverbials to the left of
the VP in English ((18-21) repeated here for convenience):
10
(41)
(42)
(43)
(44)
Jan Koster
He probably [VP saw Bill]
*He yesterday [VP saw Bill]
*He everywhere [VP saw Mary]
*He very hard [VP worked]
These facts are explained by the assumption that only probably is both in the class of
adverbials indicated as Adv1 and Adv2 in (35). The other adverbials are either exclusively
in class Adv2 (yesterday) or part of the VP (hard). As can be seen in (39), the VP ends up
in the Spec of TP, which is between Adv1 and Adv2, which, together with our
classification of adverbials, entails that only Adv1-type adverbials such as probably can
appear to the left of the VP in surface order.
In Dutch and German, in contrast, all adverbials can appear to the left of the VP.
The DPs can appear in the standard argument positions (Specs of DatP and AccP), with
only Adv1 to their left. Alternatively, they can be licensed in the Specs of PredPs, which
have both Adv1 and Adv2 to their left (see Koster 1994). These two possibilities are
confirmed by the grammaticality of both (45a) and (45b):
(45)
a.
b.
dat hij waarschijnlijk het boek gisteren las
that he probably
the book yesterday read
"that he probably read the book yesterday"
dat hij waarschijnlijk gisteren het boek las
that he probably
yesterday the book read
"that he probably read the book yesterday)
Last but not least, the observed scope facts follow straightforwardly from our assumptions. We can assume without problems that the order of adverbials in the universal
underlying structure (35) linearly corresponds with scope. So, linear scope in Dutch and
English is preserved to the extent that the underlying order is preserved.
In Dutch, all adverbials can be to the left of the verb, as partially illustrated in
(45). VP-internal material (like hard in the Dutch equivalent of he works hard) ends up in
a Spec of a Pred. This preserves the correspondence between linear order and scope.
In English, the correspondence is also preserved in cases like (46) and (47):
(46)
He probably [VP saw Bill]i yesterday ti
This sentence instantiates (39) in that the VP moved for checking ends up between an
Adv1 (probably) and an Adv2 (yesterday). Since there can be more than one Adv2
following the derived position of the VP in (39), we expect the universal correspondence
of linear order and scope to be preserved in English. This is exactly what we saw in (30),
repeated here for convenience (">" means: "has wider scope than"):
(47)
a.
b.
He saw Bill twice on his birthday [twice > on his birthday]
He saw Bill on his birthday twice [on his birthday > twice]
The Word Orders of English and Dutch
11
However, if the hypothesis of VP-movement in English is correct, we expect that the
correspondence of linear order and scope superficially seems to break down in English
when an adverbial is VP-internal. This is exactly what we observe in cases like:
(48)
He has [VP worked hard]i yesterday ti
The VP-internal adverbial hard is moved along with the VP, superficially breaking the
correspondence, but with the VP in its original position, indicated by the trace, scope
corresponds with linear order after all, as predicted by our hypothesis. In other words, we
can maintain the optimal hypothesis of Universal Grammar, namely that adverbial scope
always corresponds with linear order in underlying structure (apart from parallel
construal; see Koster 1999b).
5.
Conclusion
We started out with the observation that in Dutch subordinate clauses, the verb figures as
the center of mirror symmetry with respect to PPs. This symmetry is broken in main
clauses by the verb movement rule of Verb Second. In English, the mirror symmetry is
strikingly absent, suggesting a rule of verb movement as well. Standard Verb Second
does not work for English. However, there are strong indications that the whole VP is
moved in English for the purpose of feature checking. English VP movement for
checking the case, tense or predicate function of its constituents is a form of Pied Piping.
The same kind of checking done collectively by the VP in English is done by each of the
constituents individually in Dutch and German.
This hypothesis explains not only why the verb is not a mirror center (as in Dutch)
but also why there is a VO/OV difference between English and Dutch, why English does
not have the scrambling possibilities of Dutch and German and why English adverbials
have the anomalous order and scope properties they superficially have.
The Pied Piping parameter, in short, explains many word order facts of English
and Dutch that were hitherto unexplained. What is at least as important is that the pattern
of explanation in question was only possible under the assumptions of Kayne's antisymmetry theory, which led to an entirely new analysis of the structure of Dutch and
German based on a universal VO base and an OV order derived for feature checking. To
the extent that the analysis of this article is correct, it forms striking confirmation of the
fruitfulness of this approach.
12
Jan Koster
References
Alexiadou, Artemis. 1997. Adverb Placement: A Case Study in Antisymmetric Syntax.
Amsterdam: John Benjamins.
Barbiers, Sjef. 1995. The Syntax of Interpretation. Doctoral Dissertation, University of
Leyden, Leiden.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, Mass.: MIT Press
Cinque, Guglielmo. 1998. Adverbs and Functional Heads: A Cross-Linguistic
Perspective. New York, etc.: Oxford University Press
Dikken, Marcel den. 1996. The Minimal Links of Verb (Projection) Raising. In ed. W.
Abraham, S. Epstein, H. Thráinsson, and J.-W. Zwart, Minimal Ideas.,
Amsterdam: John Benjamins.
Jackendoff, Ray. 1972. Semantic Interpretation in Generative Grammar. Cambridge,
Mass.: MIT Press,
Kaan, Edith. 1992. A Minimalist Approach to Extraposition. MA Thesis, University of
Groningen, Groningen.
Kayne, Richard. 1994. The Antisymmetry of Syntax., Cambridge, Mass.: MIT Press
Koopman, Hilda and Anna Szabolcsi. 1998. Verbal Complexes. Ms., UCLA, Los
Angeles.
Koster, Jan. 1974. Het Werkwoord als Spiegelcentrum. Spekator 3: 601-618.
Koster, Jan. 1975. Dutch as an SOV Language. Linguistic Analysis 1: 111-136.
Koster, Jan. 1987. Domains and Dynasties. Dordrecht: Foris,
Koster, Jan. 1994 Predicate Incorporation and the Word Order of Dutch, in ed. G. Cinque,
J. Koster, J.-Y. Pollock, L. Rizzi and R. Zanuttini, Paths Towards Universal
Grammar: Studies in Honor of Richard S. Kayne. Washington DC: Georgetown
University Press.
Koster, Jan. 1999a. The Word Orders of English and Dutch: Collective vs. Individual
Checking. Ms. University of Groningen, Groningen.
Koster, Jan. 1999b. Empty Objects. Ms. University of Groningen.
Perlmutter, David. 1983. ed., Studies in Relational Grammar 1., Chicago: University of
Chicago Press.
Riemsdijk, Henk van. 1994 . Another Note on Clausal Pied-Piping, in in ed. G. Cinque, J.
Koster, J.-Y. Pollock, L. Rizzi and R. Zanuttini, Paths Towards Universal
Grammar: Studies in Honor of Richard S. Kayne. Washington DC: Georgetown
University Press.
Vanden Wyngaerd, Guido. 1989. Object Shift as an A-movement Rule. MIT Working
Papers in Linguistics 11, 256-271.
Zwart, Jan.-Wouter. 1993. Dutch Syntax: A Minimalist Approach. Doctoral Dissertation
University Groningen, Groningen.
Linguistics, University of Groningen
P.O. Box 716
9700 AS Groningen
The Netherlands
e-mail: [email protected]